Dan McLaughlin

Data Analytics for Everyone

Cognitive computing has entered the business enterprise as an important tool. The term “cognitive computing” refers to a computer’s ability to function as though it were thinking like a human. The most well-known example of this is the IBM product, Watson. Named after Thomas Watson, the founder of IBM, it is the culmination of years of research on big data and artificial intelligence and gained fame for easily trouncing two human champions on the quiz show Jeopardy.

Watson is now roaming the halls of the Opus College of Business as well. We are using a version known as Watson Analytics in our Physician Leadership College and our Executive MBA. Watson Analytics has three major cognitive components: data analytics, natural language processing and predictive capabilities. Our students enter operational data into the system and Watson initially presents a number of interesting findings in the form of graphs. Some are not useful, but many are, and the details of the analysis are easily revealed. Students can ask their own questions and Watson will present a number of answers with the most likely one presented first.

When I first started using Watson it came loaded with football data for the 2014 season. So I thought I would see how smart it was. My question was, “How come the Vikings were so bad?” Watson send me a nice graph of all the times the Vikings were intercepted during the year.

Although Watson is being used with clinical data, our students are also interested in the business applications. Our physician students complete projects with their own data. One doctor (a diabetes specialist) looked at how well his patients who were smokers complied with appointments. His working assumption was that these patients were not very compliant, but in fact Watson found that this to be false. This significantly changed how his practice operated with these patients.

Another physician supervises a large set of doctors whose primary responsibilities are for the care of patients in the hospital. A major goal for these doctors is to keep the length of stay in the hospital as short as possible, as the hospital receives a fixed payment for each patient. She took data from each hospital admission (over 1,000) and compared the length of stay for each of her doctors against the statewide average. Watson generated a histogram of the average for each doctor. Most of the physicians had averages that were close to the statewide average but three were significantly longer and one was substantially shorter. She told me “Now I have to make three difficult phone calls and give one doctor a raise.”

IBM has designed Watson to be used by the casual data miner, not the data scientist. You can try it out for free. My bet is that you will be surprised by the some of the insights a cognitive computer like Watson can provide.

Artificial intelligence, big data and cognitive applications continue to improve every day and they will be a significant force for change in many industries – especially health care.